| Wetland is one of the three major ecosystems on the earth.It is a special transitional ecosystem between water and land.As an important part of urban spatial structure,urban wetland not only plays an important role in ecological services such as environmental regulation,resource supply,disaster prevention and control,but also provides a good living environment for urban residents.Nanchang City is rich in wetland resources,but with the rapid development of its social economy,the wetland landscape pattern of Nanchang City has been constantly changing,which has a certain impact on the normal play of its ecological function.Therefore,it is of great significance to study the changing law of wetland landscape pattern and explore its evolution characteristics in Nanchang City.In recent years,the amount of data of wetland ecological monitoring has been increasing,and has gradually exceeded the traditional data.The extent to which the processing method can withstand the amount of data.The related research in the field of spatio-temporal association rules mining provides an effective solution for how to mine valuable information from massive data so as to reveal the changing rules of various wetland ecological phenomena.In this paper,the changes of wetland landscape pattern and its evolution in Nanchang from 1995 to 2016 were studied.Taking the 6 issues of Nanchang Landsat remote sensing image from 1995 to 2016 as the data source,and referring to the auxiliary data such as the land use map,the wetland information in each period of images was obtained by combining the maximumlikelihood classification with visual interpretation.The wetland information in the images was obtained by using the method of maximumlikelihood classification and visual interpretation in the light of the auxiliary data such as land-use map.According to these wetland information,the general change trend of Nanchang wetland,the analysis of wetland landscape pattern index and the driving mechanism of wetland evolution were analyzed.And using FP-growth algorithm from the wetland evolution of Nanchang city spatio-temporal events.In the traffic database,the evolution track and the change character of the wetland landscape are excavated.The results of the study are as follows:(1)In this paper,supervised classification(maximumlikelihood classification),(1)In this paper,supervised classification(maximumlikelihood classification),decision tree classification and face object classification are used to interpret the remote sensing images of Nanchang City.It is found that supervised classification(maximumlikelihood classification)has the best classification effect,and its overall classification accuracy is 83.82%.The Kappa coefficient is 0.8.The overall classification accuracy is 81.61% and the Kappa coefficient is 0.76%.The result of decision tree classification is the worst,its overall classification accuracy is 71.64%,Kappa coefficient is 0.65.After comprehensive consideration,this paper finally chooses the combination of maximumlikelihood classification and visual interpretation.Methods the remote sensing images of Nanchang City were classified to extract wetland information from Nanchang City.(2)The total wetland area of Nanchang City increased gradually in 1995 / 2001,but decreased in 2001 / 2016.The natural wetland area of Nanchang City has a steady change trend,accounting for 17.69% of the average area of Nanchang City.Lakes and swamps are the main components of the natural wetland.The area of constructed wetland in Nanchang City increased from 1995 to 2001,and decreased in 2001 / 2016.The average proportion of constructed wetland in Nanchang City was 44.22%.Paddy fields occupy a dominant position in constructed wetlands.(3)The result of patch type index analysis of wetland landscape pattern shows that the average landscape proportion index and the average maximum patch index of paddy field are both the largest,indicating that it has the highest dominance in the whole wetland landscape.Moreover,the fragmentation degree of paddy field is also the largest,and the average PD value is 2.72 / hm2,boundary complexity is also the highest.The aquaculture ponds and reservoirs are significantly affected by human activities.The spatial connectivity of lakes and paddy fields is stronger,and the spatial aggregation of lakes is the strongest.(4)The result of landscape index analysis of wetland landscape pattern showed that the fragmentation degree of wetland landscape decreased from 1995 to 2001,and the influence of the largest patch increased from 1995 to 2001.During the period from 1995 to 2001,the fragmentation of wetland landscape increased,and the maximum patch influence weakened.The overall wetland landscape is affected by human activities first and then intensifies and then weakens,and the connectivity of all types of wetland landscape is contrary to that of other types of wetland landscape.The balance of landscape distribution increased gradually after 2001,and the aggregation degree of all kinds of wetland landscape showed the trend of increasing first,then decreasing,then strengthening and then weakening.(5)Based on the analysis of driving mechanism of wetland landscape by principal component analysis,it is found that economic development,population change,secondary industrial structure,precipitation and air temperature are the main driving forces of wetland landscape change in Nanchang.(6)A total of 9386 strong association rules were excavated from the spatio-temporal database of wetland evolution trajectory in Nanchang.Among them,941 association rules accord with the evaluation criteria of correctness,timing and evolution.Based on the statistical analysis of these interesting association rules,it is found that in 1995 / 2016,wetland type land mainly evolved to construction land and swamp land types.The number of wetlands transformed into construction land is the most,with a total of 285 tracks of the evolution of construction land,indicating that the urbanization construction and precipitation in Nanchang have the most significant impact on its wetland evolution.With the evolution of wetlands,the association rule dimension of the If the confidence level is increased,the average confidence level will also increase.In the association rules of five-dimensional wetland evolution trajectory,the number of tracks turning to construction land is the most,and paddy field is the main type of evolution. |
PDF Full Text Request